Re: LUN field in iSCSI Data PDU

[julian_satran@il.ibm.com]

It makes the tag allocation at the target so much easier. Tags are now a LU
local issue.
You don't have to take this path but if you are building a very large
target you may want to.
Anyhow I did no hear any objection to it when I first suggested it (a
century ago) as a way
to "localize" target Tags and it comes at (almost) no cost.

Julo

Matt Wakeley <matt_wakeley@agilent.com> on 21/11/2000 23:01:39

Please respond to Matt Wakeley <matt_wakeley@agilent.com>

To:   IPS Reflector <ips@ece.cmu.edu>
cc:
Subject:  Re: LUN field in iSCSI Data PDU




julian_satran@il.ibm.com wrote:
>
> Vengat,
>
> The LUN field is mandated only if the Target Tag is valid (i.e., the data
> is sent in response to a R2T).
> This way you can get to "decentralize" the target activity as the tags
can
> be issued "locally" by the LUs.

This sounds like a new "feature" that isn't required.  Fibre Channel
doesn't
require LUN in any phase of it's transfer except the command - why must
iSCSI
require LUN in data transfers?

-Matt

>
> Julo
>
> "Venkat Rangan" <venkat@rhapsodynetworks.com> on 20/11/2000 20:38:23
>
> Please respond to "Venkat Rangan" <venkat@rhapsodynetworks.com>
>
> To:   "IP Storage Working Group" <ips@ece.cmu.edu>
> cc:
> Subject:  LUN field in iSCSI Data PDU
>
> Section 2.8 specifies that the SCSI Data Header contains "LUN or Reserved
> (0)" at offset 8-16 within the header. I am not clear under what
> circumstandes it contains a LUN. Note that the LUN info was already sent
in
> an earlier SCSI CMD PDU. If a LUN is specified in SCSI Data PDU as well,
> what is its relationship to the one sent in SCSI CMD?
>
> If we look at Fibre Channel FCP_DATA framing, it does not include a LUN,
> since the FCP_CMD has established the LUN for the exchange already.
>
> Venkat Rangan
> Rhapsody Networks Inc.
> www.rhapsodynetworks.com